3DMPR - a robust morphological approach for applying phase retrieval in proximity to highly attenuating objects in...

Area of Science:

• Medical Imaging
• Computational Imaging
• Materials Science

Background:

• X-ray imaging with computed tomography offers detailed 3D sample visualization.
• Propagation-based phase contrast enhances data quality for low-attenuating samples, reducing radiation dose.
• Conventional phase retrieval struggles with multi-material samples, causing over-blurring and data loss.

Purpose of the Study:

• To develop an efficient, non-iterative method for strong phase retrieval in 3D X-ray imaging.
• To preserve sharp boundaries of all materials in multi-material samples.
• To improve signal-to-noise ratio (SNR) and reduce blurring artifacts in phase-contrast imaging.

Main Methods:

• Introduced a novel technique, 3D phase retrieval with morphological operations (3DMPR).
• Combined 3D phase retrieval with morphological operations to prevent over-blurring.
• Tested 3DMPR on synchrotron X-ray phase contrast images of a rabbit kitten brain and an aluminum-water phantom.

Main Results:

• 3DMPR achieved a 6.8-fold SNR improvement for brain tissue without over-smoothing.
• Tested on a phantom, 3DMPR provided a four-fold SNR boost.
• 3DMPR preserved boundary spatial resolution at 54 ± 1 µm, compared to 108 ± 2 µm with conventional methods.

Conclusions:

• 3DMPR is an efficient, non-AI, non-iterative method for superior phase retrieval in multi-material samples.
• The technique significantly enhances image quality and quantitative accuracy.
• 3DMPR opens possibilities for dose reduction in clinical X-ray imaging applications.